Shared publicly  - 
 
My colleague Jon Bruner published a short piece on Radar about using sensors to better manage machine tools and other machinery today.

http://radar.oreilly.com/2012/10/listening-for-tired-machinery.html

The piece brought back some memories. I was in the U.S. Navy for about ten years and in our world the Preventive Maintenance System was gospel. Every system and piece of equipment on board ship had a PMS schedule and, of course, it was all based on best guesses of wear and tear at normal operating cycles. In those days, at least for most systems, we didn't modify the  schedule for real world use.

But in 1992 or so we started experimenting with hand held vibration analyzers that we would periodically attach to hard points on machinery. By monitoring vibration the analyzer could make some predictions about bearing wear and failure and we could start to modify our schedule where appropriate.

In the near term the kinds of research Jon points to will see more and more machinery instrumented. But to me the interesting part won't just come from instrumenting individual pieces but in their combination. An instrumented device can run (relatively) static algorithms that predict when it requires maintenance, but an entire shop floor of instrumented equipment can interact with scheduling and other systems to optimize maintenance schedules to maximize plant throughput. This is just at the beginning.
4
Jon Bruner's profile photo
 
Great point--the opportunity for optimization grows as the scale of the system being encompassed in the optimization gets larger.
Add a comment...